Dr. Andres Perez - PRRS Epidemiology: Best Principles of Control at a Regional Level

Information about Dr. Andres Perez - PRRS Epidemiology: Best Principles of Control at a...

Published on January 17, 2016

Author: trufflemedia

Source: slideshare.net

Content

1. PRRS Epidemiology Best principles of control at a regional level Andres Perez, DVM, PhD Endowed Chair of Global Animal Health and Food Safety University of Minnesota [email protected] Chicago, December 2015

2. Swine Group, University of Minnesota Mike Murtaugh: Xiong Wang Bob Morrison: Carl Betlach Andres Perez: Pablo Valdes, Moh Alkhamis, Julio Alvarez, Kim VanderWaal Team work Project participants Swine Health Monitoring Program: 4 anonymous participants Regional Control Program N212 (RCP N212): Dave Wright (coordinator) and anonymous program participants Sponsors Swine Health Information Center National Pork Board University of Minnesota Population Systems and MnDrive programs Boehringer Ingelheim

3. A. Vaccination or exposure to live-virus B. Elimination +sows (test-removal) C. On-site/off-site testng D. Cleaning & disinfections of trucks E. Aerosol filtration F. Improve biosecurity Control at the farm level “Strategies to control at the farm-level perform reasonably well, but we still need to understand how to control the disease at the regional level” (Polson, Mondaca, Cano 2006)

4. To develop methodological frameworks to: 1. Evaluate progress of Regional Control Programs, RCPs (study 1) 2. Identify the emergence and monitor the spread of new PRRSv strains (study 2) Objectives

5. Study 1: Methodological framework to evaluate progress of RCPs Pablo Valdes-Donoso, Lovell S. Jarvis, Dave Wright, Julio Alvarez, Andres M. Perez. Measuring progress on the control of porcine reproductive and respiratory syndrome (PRRS) at a regional level: the Minnesota N212 regional control project (RCP) as a working example. PLOS One. Submitted. Objectives: To evaluate: 1. Demographics of enrollment 2. Demographics of active participation (sharing PRRS status) 3. Trend and spatial distribution of incident cases

6. Data: N212 6 A. Geographical location B. Date of enrollment C. PRRS status (weekly) D. Type of farms 1. Farm level 2. Longitudinal (weekly) 3. June 2012 – June 2014 (2 years)

7. 1. Demographics of enrollment - Proportion of farms (with sows, SS and without sows, NSS) enrolled in the RCP-N212 - ANOVA - Null hypothesis: the proportion of SS and NS farms enrolled in the RCP-N212 was constant through time Statistical analysis (1/3)

8. 2. Demographics of active participation - GLME model for binary response - Response variable: Sharing PRRS status (Y=1, N=0) - Effects: - Fixed Effects: Time, farm type - Random Effects: Site, county - Spatial and temporal analysis of sharing PRRS: - Normal Scan Statistic Test (SaTScan) Statistical analysis (2/3)

9. 3. Trend and distribution of PRRS incidence Statistical analysis (3/3) -GLME model for binary response - Response variable: PRRS status (Pos=1, Neg=0) - Effects: - Fixed Effects: Time, farm type, probability of reporting PRRS, Farm density in the county, Proportion of vaccinated farms in the county - Random Effects: Site, county - Time-space correlation of pairs of incident cases

10. Results 1/3 Number of farms enrolled and geographical coverage increased Ratio: 1/3 sites SS/NSS Proportion of SS and NSS, did not change over time (p>0.05) 1. Demographics of enrollment

11. Results 2/3 2. Demographics of active participation (sharing) Participation increased significantly (p<0.001) NSS less prone to report than SS (but improved lately!) Variability of participation higher between than within counties (RE)

12. Results 2/3 2. Demographics of active participation (sharing) Low probability of sharing PRRS data during first half (Jul12-Jun13) High probability of sharing PRRS cluster during second half (Jul13-Jun14) * Counties delimited have sites enrolled in RCP-N21 ** Gradient per county indicates number of sites

13. Results 3/3 3. PRRS trend A significant (p<0.001) monthly decrease of PRRS incidence Probability of sharing PRRS status negatively associated with PRRS incidence Density of large and medium-sized sites in county positively related with PRRS incidence Proportion of vaccinated farms not associated with the disease

14. Results 3/3 3. PRRS trend Spatial and temporal distribution of pairs of incident cases

15. Results 3/3 3. PRRS trend Spatial and temporal distribution of pairs of incident cases • Temporal windows of < 3 weeks • Spatial windows of <3 kilometers 1 5

16. 1. Three-stage systematic approach to evaluate the evolution of RCPs (enrollment, sharing, incidence) 2. Farmers’ enrollment not necessarily a good estimate of participation (may not share information on disease status) 3. Sharing information has increased, but NSS (~77%) less prone to report (although it improved) 4. Sharing information among producers negatively correlated with disease incidence 5. PRRS incidence has decreased, but spatial and temporal aggregations remained over the study period Discussion

17. 1. To apply this systematic approach to other RCPs (anybody interested????) 2. To evaluate incentives and deterrence of participation in RCPs and collaborative strategies to control PRRS at a regional level Future plans

18. Objective: To propose a methodological approach to monitor emergence and spread of PRRSv strains Study 2: Identify the emergence and monitor the spread of new PRRSv strains Moh Alkhamis, Andres Perez, Michael Murtaugh, Xiong Wang, Bob Morrison. Applications of Bayesian Phylodynamic Methods in a Recent U.S. Porcine Reproductive and Respiratory Syndrome Virus Outbreak. Frontiers in Veterinary Science, submitted.

19. Paton et al. 2005. Selection of foot and mouth disease vaccine strains- a rev Introduction Suppose that we have sequences from 8 farms (1 per farm) Each square represents one nucleotide of the sequence

20. Paton et al. 2005. Selection of foot and mouth disease vaccine strains- a rev Introduction Alignment

21. Paton et al. 2005. Selection of foot and mouth disease vaccine strains- a rev Matrix of (nt) differences 1 2 3 4 5 6 7 2 2 3 1 2 4 5 4 4 5 2 1 1 3 6 4 3 3 2 2 7 3 1 3 4 2 4 8 6 5 5 1 5 3 5

22. Paton et al. 2005. Selection of foot and mouth disease vaccine strains- a rev (nt) Genetic distance 1 2 3 4 5 6 7 2 0.333 3 0.167 0.333 4 0.833 0.667 0.667 5 0.333 0.167 0.167 6 0.667 0.5 0.5 0.333 0.333 7 0.5 0.167 0.5 0.667 0.333 0.667 8 1 0.833 0.833 0.167 0.833 0.5 0.833 Genetic distance (GD) GD = (nt, aa) differences / (nt, aa) compared Probability of finding one (nt, aa) difference when two strains are compared

23. Paton et al. 2005. Selection of foot and mouth disease vaccine strains- a rev Phylogenetic trees 1) Because sequences theoretically split into two descendant sequences, phylogenetic trees are typically assumed to be bifurcating 2) Topology: branching pattern of a tree 3) Taxa: any kind of taxonomic units (eg.: virus sequences) Unrooted tree Rooted tree

24. Paton et al. 2005. Selection of foot and mouth disease vaccine strains- a rev Phylogenetic trees If 4 taxa are being compared, there are 15 and 3 possible rooted and unrooted tree topologies, respectively From Nei and Kumar, 2000

25. Paton et al. 2005. Selection of foot and mouth disease vaccine strains- a rev Phylogenetic trees When 10 taxa are being compared, the number of possible rooted topologies is 34,459,425 and the number of possible bifurcating unrooted topologies is 2,027,025 Only one of them is the true topology Reconstructed or inferred trees: trees built from observed sequences Traditional methods to infer a tree (distance, parsimony, likelihood) ignore the associated epidemiological information (space, time of identification, associated data)

26. Correlation of alternative metrics of PRRSv relatedness, namely, difference (%) in the number of nucleotides (X axis) and considering phylogenetic evolution (Y axis). Correlation is quite bad

27. Phylogenetic methods • We can use phylogenetic methods to infer: – Most likely time of emergence of the strain; – Past probability of spread between systems and between type of farms – Population (viral) size

28. Data • 6,774 ORF 5 sequences • January 1998 – April 2015 • 5 independent systems (coded A…E) • Type of farm (sow, growing pig farms) • Selected 1-7-4 PRRSv using a maximum likelihood algorithm

29. Results Some (n= 288) related sequences (based on ML) collected between September 2003-March 2015 Most (n=241) are 1-7-4 sequences obtained between Jan 2014 and Mar 2015

30. Results Increase in population size in 2014

31. Results History of spread between systems

32. Results History of spread between types of farms

33. Dynamic representation

34. Future plans • Continue enrollment of systems • Standardization and optimization of protocols for data collection and sharing • Implementation of analytical tools into IT systems • Provide near real time interpretation of results • Contribute to PRRS control at a national scale

35. Thank you

Related presentations


Other presentations created by trufflemedia

2009 Iowa Pork Congress Seminars
05. 02. 2009
0 views

2009 Iowa Pork Congress Seminars

2009 Pig Ski Agenda
18. 02. 2009
0 views

2009 Pig Ski Agenda

Law
Law
Dean Black - Beef Trade Missions
17. 04. 2015
0 views

Dean Black - Beef Trade Missions